Dante audio protcol over IP

[sklankowski]

I have a unique application with operation in large park setting, but I have been trying to find a better method of creating a single distributed sound throughout various display areas without using FM broadcasting and reamplify.  I was told that if I was to use the audio from a remote Pi that it would likely be stuttering due to timing corrections.  I have been looking at plugins that stream audio, but that would a device to process that stream.  lately, I  have been looking at Dante audio network equipment, to utilize a USB to IP adapter from FPP and cast it across a network (will have hard line IP to large props).  I am trying to search for code/scripts to enable the FPP to send out a Dante protocol signal, but I could experiment with the USB and physical adapters. 
Curious if someone could help on this pursuit to add feature. Found a project that is closely related to my desire, but lack more indepth coding knowledge.      https://github.com/jsharkey/wycliffe

[algerdes]

How big is "big"?  Distance between your outer and central units would help.

[sklankowski]

90 meters (300ft) for now, from FPP master to walk through tunnel slave.   With future progress moving South in coming years.  I figured since I will have to build a robust network to connect elements, I would incorporate the audio in the network/separate network.   

[algerdes]

2 things to keep in mind:

* There is a real possibility of the stutter that you have been told when using remotes to play audio.
* Even if you should use a different propagation method, you will still need to amplify whatever signal to push appropriate speakers. 

There are several different pro level devices for transmitting audio from point a to point b over the distances you are showing, but they are far more costly than an FM transmitter and receivers.

If you can run cable, 70v (or 100v) speaker feeds can go a very long way.  I've set speakers every 100 feet down a 1000 foot run for some street fairs and it worked great.  Takes a special amplifier with 70v outputs and speakers with appropriate transformers.

[jcross]

2 things to keep in mind:

* There is a real possibility of the stutter that you have been told when using remotes to play audio.
* Even if you should use a different propagation method, you will still need to amplify whatever signal to push appropriate speakers. 

There are several different pro level devices for transmitting audio from point a to point b over the distances you are showing, but they are far more costly than an FM transmitter and receivers.

If you can run cable, 70v (or 100v) speaker feeds can go a very long way.  I've set speakers every 100 feet down a 1000 foot run for some street fairs and it worked great.  Takes a special amplifier with 70v outputs and speakers with appropriate transformers.

+1 on 70/100V.


In my past life we used to provide paging to a 16 acre fair ground using a distributed 100V system with large format horns.


I have also done a setup with an FM transmitter and put receivers with amps/speakers all over the place.
I also did one show (an outdoor car show) where we broadcast the main stage over FM and just turned on all the car radios. That was really simple!


From the info provided I would suggest that your cheapest method would be a single FM transmitter and multiple cheap FM receivers with powered speakers (computer speakers work if people are close enough).
For the haunted house I do I picked up several cheap paging horns and car speakers.  I found some 12VDC marine (boat) amplifiers and put those in each room.  In this case I provide audio from the PI (not running FPP as each room is it's own world) but it is just an audio input so a $10 FM receiver would be a suitable input for it.

[sklankowski]

The primary reason I am avoiding FM transmission is that I have a significant number of commercial FM stations operating at 100 kW or more due to the geolocial features in my area (Bluff Country - La Crosse WI).  And I am currently broadcasting on a new station after changing from another this year due to additional commercial use of the spectrum in the area.  This is why I am looking to other options.
I have used 70V/100V amplifiers before, and yes it could work in this application.  but then I have to have this HV power audio signal run to each prop/zone that we want the same mix to broadcast. That could be cumbersome. 

I am liking this Dante protocol as I can tomorrow use one of their USB output devices to encode audio and cast it over the same network cable or on a second network.  In the near future, the group I am with wants the entire 1500ft park to blink with music.  There will be a concurrent FM broadcast for the cars, but for the walkers to enjoy, I figure I could just hop off any one of the Falcon network ports at a display with another audio output device and have the clean audio signal, no stuttering. 

It is the latency question that induces the syncing issue of audio from a falcon device that i am worrying about using these adapters and my desiring to minimize at each of these areas.  The display areas are currently separated enough that people may not notice if the unit is behind the other area by 300ms, but the sync between lights and audio could be noticeable and music stuttering would definitely be a joykill. 

I was posting this to start a conversation to see if the FPP's could output both the E1.31 signal and possibly this Dante signal on the same ethernet port, someday, reducing the latency of using the USB to IP adapter. 
https://www.audinate.com/content/dante-avio-usb-adp-usb-au-2x2-firmware-v4153

As for cost, I see $500 on these adapters now would become equivalent to installing multiple FM receivers, powered speakers, or amplifier and 2000ft of 14 awg audio-signal/speaker wire, not in a straight line. I would  already have the cat5 there and switchgear. 

[CaptainMurdoch]

Are you wanting to replace the source/transmission side of the Dante system or the destination/receive side?

I took a look at https://www.audinate.com/products/devices/dante-avio#usb and it looks like the USB transmitter is only $129 and an analog receivers are $129 each.  So, $360 for a pair and then tack on an additional $129 per additional receiver.

[dkulp]

If the USB adapter thing implements standard USB audio protocols, it likely would "just work" out of the box right now. That said, I'm not going to pay the $250 to give it a try. 

[sklankowski]

Are you wanting to replace the source/transmission side of the Dante system or the destination/receive side?

I took a look at https://www.audinate.com/products/devices/dante-avio#usb and it looks like the USB transmitter is only $129 and an analog receivers are $129 each.  So, $360 for a pair and then tack on an additional $129 per additional receiver.

I am looking at the source/transmission side.

[CaptainMurdoch]

I am looking at the source/transmission side.

If you only have once source then it seems like you would only need to use the AVIO USB sound card adapter which should work with FPP since it is supported in Linux according to the page I linked.  I haven't found any specific examples online, but as Dan said, if it emulates a standard USB sound device then it should work out of the box with our FPP images.

[sklankowski]

If you only have once source then it seems like you would only need to use the AVIO USB sound card adapter which should work with FPP since it is supported in Linux according to the page I linked.  I haven't found any specific examples online, but as Dan said, if it emulates a standard USB sound device then it should work out of the box with our FPP images.

I will try it out soon. Thank you for the assurance.  I was looking into the use of POE powered speakers that have Dante built in.  Costly now but in the future maybe. 

[algerdes]

I'm very curious about several things:
1.  How much latency does it (the dante pair(s)) inject in the audio stream?
2.  How much additional latency is encountered going through all the switches?
3.  How much bandwidth does it take up? 
4.  Can it coexist with the signals from the FPP to the controllers?  Both with sync signals between FPPs or full control signals between the FPP and directly to controllers such as the Falcons.

Edit - I have watched the first 7 videos that Audinate produced about Dante.  At least one of my list have been answered:
1.  Latency is usually within 10ms.  According to them, it is as low as 1ms.  This will be nice to see.

Note that they make it very clear that they should only be used on a wired network.  Apparently (and this is from their videos, and not any hands on) wireless networks add too much latency for their interactive program to be of any use.  (Partially answering #2)

One of the things that stands out for me is the ability to send Dante packets to multiple receivers at the same time.  Nice!  Multiple channels across one Ethernet connection is also a plus.

[sklankowski]

I'm very curious about several things:
1.  How much latency does it (the dante pair(s)) inject in the audio stream?
2.  How much additional latency is encountered going through all the switches?
3.  How much bandwidth does it take up? 
4.  Can it coexist with the signals from the FPP to the controllers?  Both with sync signals between FPPs or full control signals between the FPP and directly to controllers such as the Falcons.

Edit - I have watched the first 7 videos that Audinate produced about Dante.  At least one of my list have been answered:
1.  Latency is usually within 10ms.  According to them, it is as low as 1ms.  This will be nice to see.

Note that they make it very clear that they should only be used on a wired network.  Apparently (and this is from their videos, and not any hands on) wireless networks add too much latency for their interactive program to be of any use.  (Partially answering #2)

One of the things that stands out for me is the ability to send Dante packets to multiple receivers at the same time.  Nice!  Multiple channels across one Ethernet connection is also a plus.

Your questions are exactly what I will determine this Spring when I assemble the network backbone structure in the park.  At present, my friend and I have been researching the latency within switch gear, searching spec online, showing <15ms. 
Yes, I do believe WiFi would introduce significant more latency and buffering, but how much and will it be noticeable, may depend on gear.   
The main reason I am researching this protocol/setup is inline with your last comment, the ability to send multiple channels down a stream and selectively choose which channels land at specific speakers. Surround sound??  PoE cube speakers??

Question to answer: how does the internal clocks correct or push this effect to where it will be noticeable?
Hypothesis 1:  If E1.31 and Dante data are inputted into same switch at identical time, passed across a network backbone, and outputted from another switch to corresponding components, then both will have the same latency factor corresponding those network node/star points.  Factors: how network is built.   

It should be able to coexist if utilizing the same network. 
Proposed Answer to #3
Bandwidth = (Sample rate) x (Bit depth)
Example: 48kHz sample rate, 24-bit depth
48,000 x 24 = 1.152 mbits/sec per channel
64 channels of audio at 48kHz/24-bit
48,000 x 24 x 64 = 74 mbits/sec   (9MB/sec!;  This may be just the audio signal, not the overhead) Proposed Answer to #4:  The last time I did any estimations with sACN/E1.31; you can run about 35 universes ~10 mbps, 350 on ~100 mbps, ~3500 on 1 gbps.   I also know that it requires some overhead.  Gigabit switches!
     
Reference:  https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=2ahUKEwiFlrnFktjgAhUTtIMKHeEUCeoQFjAAegQIChAC&url=https%3A%2F%2Fwww.audinate.com%2Fsites%2Fdefault%2Ffiles%2FPDF%2Fdante-certification-level-1-introduction-audio-networking-audinate-pres.pdf&usg=AOvVaw2w3Sidk-jCP8P1QQv_Pt0x 
 
 
I have purchased the USB module and the XLR output module.  I was also looking at the individual Adapter Module Boards to use other connectors (output to 1/4" or 1/8") directly. 

[algerdes]

I'm very curious as to why the folks at Audinate are pretty animate as to NOT using WiFi. 

Timing between each device is possibly the reason.  Each device runs its own clock, with a signal from the "master" telling the other clocks where it is at specific intervals. 

The other thing is the automatic drop-out/re-sync when a device is disconnected and then reconnected.  I wonder if the lag introduced as part of the WiFi interaction would be enough to make it believe that a device is not there - even if it is.

In any case, thank you for bringing this all to my attention.  I am "old school" as I have been doing sound for events for over 5 decades.  I still believe in "if it don't move, wire it!", but am always open to different ideas.

[sklankowski]

I'm very curious as to why the folks at Audinate are pretty animate as to NOT using WiFi. 

Timing between each device is possibly the reason.  Each device runs its own clock, with a signal from the "master" telling the other clocks where it is at specific intervals. 

The other thing is the automatic drop-out/re-sync when a device is disconnected and then reconnected.  I wonder if the lag introduced as part of the WiFi interaction would be enough to make it believe that a device is not there - even if it is.

In any case, thank you for bringing this all to my attention.  I am "old school" as I have been doing sound for events for over 5 decades.  I still believe in "if it don't move, wire it!", but am always open to different ideas.

You are most welcome.  I like the interoperability of adding different devices anywhere along a network to have audio outputs, being that this park I am in has a large aspect ratio.  This also allows the connectivity of many display elements that are separated by tens of meters.  plus I like learning about the new technology.